【作者】 王伟；

【导师】 许鑫；

【作者基本信息】 华东师范大学 ， 情报学， 2011， 硕士

【摘要】 随着网络的蓬勃发展,互联网已经成为人们发布和获取信息的重要渠道,网络信息越来越被人们所关注,因此对网络信息的热点事件发现与分析是十分必要的。从互联网的特性来看,网络信息来源众多,随机性强,信息发布者的观点和角度各不相同,仅仅靠传统的经验判断无法帮助用户了解网络信息主要热点事件和某个热点事件的主要方面。因此,需要采用一定的技术与方法对网络信息进行自动处理,用于快速准确的发现网络信息的热点事件,同时可以对热点事件进行一定的分析研究。本文以互联网网页信息为研究对象,利用信息采集技术、聚类技术等为网络信息的发现与分析提供了一套有效的解决方案,使用户能够清晰的了解当前社会的热点事件,并在一定程度上对热点事件进行规律分析。首先,本文介绍了网络信息热点事件发现与分析的背景、国内外研究现状和热点发现与分析中需要用到的一些相关关键技术。其次,着重就本文提出的两个创新点进行了介绍,即网络信息采集策略的改进和聚类算法的改进。通过以上改进,在一定程度上提高了信息采集效率和热点发现的效果。接着,本文就采集的结果提出了关于热点事件趋势发展的一些模型,来对网络信息热点进行分析与预测。最后,本文以创业板上市公司为例,就热点发现与分析进行了案例实验,实验证明本文提出的一些思路取得了一定的效果。热点发现与分析技术在国内的研究还比较落后,存在着大量的问题有待改进,这也意味着该研究有着巨大的提升空间。最后,本文对已做的工作进行了总结并对未来的研究进行了展望。更多还原

【Abstract】 With the vigorous development of the network, the Internet has become an important way to issue and access information.Network information has been growing concern. So it is necessary to discovery and analyze hot event information on the network.From the characteristics of the Internet,there are many sources of the information,Network information is very random.the angle and view of the information publishers are varies.We can not help users to understand all of the hot information or all aspect of some hot information from just judging by the experience of traditional. Therefore, we need to adopt a certain of technology and methods to automatic process network information and find the hot events quickly and accurately from the network information. At the same time, we also can do some prediction and analysis.In this paper, my research object is the Internet web page information. We provides an effective solution that make users clearly understand the current hot issues of society and predict the hot events to some extent by the using of information collection technology, clustering technology and so on.First, this paper describes the background, research status, and some of the key technologies of the network information discovery and analysis. Secondly, this paper focuses on two innovations that is the improving of the network information acquisition strategies and the improving of clustering algorithm. Through the above improvement, the effect of hot events detection and the efficiency of information collection have been improved in a certain extent. Then, this paper proposed some models on the development of the hot events to analyze and predict the network information hot events. Finally, we take the GEM listed companies as an example, it shows that these improvements achieved a certain results.Hot event discovery and analysis in domestic research is still relatively backward, there are lots of issues to resolve, which means the research has great space for improvement. Finally, the work of the paper and future research are discussed.更多还原